Our projection is that the vast majority of quantum computing usage, at least for the rest of this decade, will be through cloud-based quantum computers located at an external vendor’s remote site rather than through on-premises installations at the end user’s site. There are a number of logistic challenges with on-premises quantum installations that can make these difficult. These include the initial purchase cost of the computer, requirements for frequent hardware and software upgrades from the vendor, reliability issues, large maintenance and calibration needs, availability of spare parts, complex facilities requirements, and a desire to get the high utilization rates out of a very expensive machine.
But cloud installations have their downside too. The largest being the potential security and privacy concerns a user could have with a machine that is not entirely within their control. If there is a bad actor at the cloud vendor, they could potentially steal the data or the quantum algorithms or otherwise do nefarious things that could cause a lot of problems for a remote user. This issue is a concern in both classical and quantum computing and is one of the reasons people are sometimes reluctant to use cloud-based computers, even in the classical world.
One area of research in classical computing is called homomorphic encryption. This is a method that allows a user to encrypt their data at their own site before sending it for processing to a powerful computer in the cloud. The cloud processor can implement an algorithm even while the data is still encrypted and then send the encrypted result back to the end user. The end user can decrypt the result and get the answer they wanted. An unintended person at the cloud vendor would not be able to determine the input data or the output data because the data is encrypted, so the information stays private for the end user. Someone at the cloud vendor, though, could still get their hands on the program that was used, and this could still be a potential security issue. Homomorphic encryption is not able to hide the program or the algorithms that the classical could machine used in the computation. Homomorphic encryption is a step in the right direction, but it may not be a complete solution for someone who is using an untrusted computer at a remote site.
Blind quantum computing goes a step further and can not only hide the input and output data, but also the program used at the remote cloud-based quantum computer. It does require a quantum internet connection to the quantum machine in the cloud, but potentially this approach can provide the same level of security and privacy as having the quantum machine on-premises, without the associated on-premises logistical issues. A review of blind quantum computing and related protocols can be seen in an article available on the Nature website here.
One of the companies with considerable expertise in blind quantum computing is Horizon Quantum Computing located in Singapore. Their CEO, Joe Fitzsimons first invented this protocol in 2008 with Anne Broadbent and Elham Kashefi, and also their chief scientist, Si-Hui Tan, has also worked on a number of closely related results on quantum homomorphic encryption.
Recently, Horizon has been selected to become a node on Singapore’s National Quantum-Safe Network (NQSN) just announced by the National Research Foundation, Singapore, and the National University of Singapore (NUS). This network will initially include 10 network nodes to be installed across Singapore connected to fiber, including two at NUS, two at the Nanyang Technological University, Singapore (NTU Singapore), a node at Horizon, and others at government and private company premises. This quantum network program will be funded with $8.5 million Singapore dollars ($6.3M USD) over a three-year period. The main goal of this effort will be to develop a means to provide robust cybersecurity for critical infrastructure including communication systems for governments, critical infrastructure such as energy grids, and companies handling sensitive data in areas such as healthcare and finance. Key areas for research will include implementation of Quantum Key Distribution (QKD) and Post Quantum Cryptography (PQC) for full security against quantum computer attacks of general data transmissions. And Horizon will also be participating by demonstrating the key building blocks for secure quantum computing, including blind quantum computing as one of the areas of research envisioned under their Memorandum of Understanding (MoU with NUS.
For additional information about Singapore’s plans to build a national quantum-safe network and Horizon’s participation in this program, you can view a press release from NUS announcing the program here and another press release from Horizon announcing their participation in the program here.
February 19, 2022